I am using an old paper, published in 1986. It is in the form of a "Letter to Nature" (which is a scientific journal). It doesn't appear to be typeset in TeX, at least not TeX as we know it today. Furthermore, it is a scanned copy, so it is just a image, embedded as a PDF. The lack of multiline equations, and highlight-able text is getting to me.

I'm considering re-typesetting it myself by hand in LaTeX (it is quite short) for my own benefit and understanding.

The paper is still being cited today, and is used in university courses to teach the subject. It seems a waste to have gone to the effort to retype it neater, and then just leave it in my draws. If I re-typeset it, what can/should I do with the new version?

Can I host it online myself (giving full credit to the authors)? Should I send a copy to authors, for them to do with as they will? Are there issues with the fact that it has been published in Nature? Perhaps they have some copyright on it?

  • 12
    It's almost certainly copyrighted, probably by the journal, otherwise by the authors. If you distribute your copy without their permission, legal action is quite possible. Jan 15 '14 at 2:41
  • 2
    One thing to be aware of is the possibility of introducing errors in a retypeset version. We all think we never make mistakes (!) but... errors happen. If retypesetting became common practice one would fear a propagation of errors. Remember that published versions have been carefully proofread by the author, which is likely not to happen for a retypeset version. Also, the scanned version on your link is admittedly of rather poor quality but new copies/scans can be made from a copy of the journal, which I suspect would be just as useful.
    – A.G.
    Jan 15 '14 at 23:58
  • This is most certainly not typeset in TeX (which hasn't changed much since 1982 and not since 1990) but something else. Jan 16 '14 at 7:59
  • I just downloaded a clean legal copy directly from nature and ran Adobe's built in OCR on it. It is not perfect, but it may be good enough.
    – StrongBad
    Jan 17 '14 at 16:04
  • 2
    Seems to be hosted on one of the authors' website as well Learning representations by back-propagating errors. Jan 21 '14 at 15:28

The copyright will be owned by Nature, so it would be illegitimate (and, as @AnonymousMathematician says, somewhat rude) for you to distribute it yourself.

However sending a re-typeset version to the original authors isn't distribution, but simply correspondence. If they want to re-distribute this, they're in a better place to do so than you are. This may not have occurred to them, and your action in suggesting they do so, will be at the very least encouragingly flattering!

I doubt Nature would be opposed in principle. Their author licence policy says that ‘[Nature Publishing Group] actively supports the self-archiving process, and continues to work with authors, readers, subscribers and site-license holders to develop its policy.’ That's a fairly vague remark, of course, but many universities and funders now mandate 'self-archiving' papers, and journals have no option but to go along with this; therefore the authors may have already have a way of distributing this in a way which is low-hassle to them, and unobjectionable to the journal(s).

Detail: Authors aren't, typically, allowed to distribute the publisher's PDF version of a paper, but are increasingly allowed to 'self-archive' and distribute the 'post-refereeing authors' version' (that is, the same text as in the published version, but typeset by the author).

So, get in touch with the authors. It sounds like you'll be doing the community a favour.


The copyright is almost certainly owned by Nature. I don't know what sort of rights the authors retained in Nature's copyright agreement from 1986, or how to find out other than by asking Nature or the authors. My guess is that Nature will not want you to distribute the paper, and the authors may not even have a copy of the agreement or remember what was in it, but who knows. Nowadays Nature allows authors to post their own version of the paper on their own web site or institutional repository six months after publication. This clause couldn't have been in the 1986 agreement (there was no web back then), but maybe Nature would agree to it retroactively. If so, then the authors could legally distribute your retyped version. It's hard to say whether Nature will agree or whether the authors will want to bother with this, but it could be worth asking.

Of course you could retype and distribute it without permission. Even aside from the illegality, it seems a little rude to distribute a retyped version without checking with the authors. For example, if you introduce typos, readers may not know who to blame, you or the authors. On the other hand, the authors probably can't legally authorize you to distribute it, so it's an awkward situation.

It's too bad our laws and customs make it difficult to arrange things like this. When someone does retype a famous paper (such as Shannon's 1948 paper introducing information theory), it can be a really useful service for the community. I hope you are able to find a way to do this with Nature and the authors' permission.

  • Would I even be able to send a copy to the authors? Given that I would be distributing something that I don't own the copyright to? Jan 15 '14 at 2:57
  • You're right that there is a technicality here: Nature could agree to let the authors distribute it but not grant you permission to give it to the authors. On the other hand, if Nature is willing to do anything here, then I can't imagine they will care whether you retype it or the authors do. If you end up asking them, it's safest to be explicit about what you're asking for (i.e., would they allow the authors to distribute a version you would retype?). Jan 15 '14 at 3:13

While both answers above are correct in that you'd be violating copyright by typesetting the paper and distributing it, it sounds like your intent here is to create a document for your understanding, and that duplicating the paper is the way you're proposing to do that.

Another way is to think of how you'd prepare lecture notes based on this paper. It's very common to see lecture notes that go into detail on a single paper. Now you can't reproduce figures/tables from the original work without permission, but you can definitely explain the paper using your own understanding it, and referring readers to the original source for any specific figures/tables/results. Since this particular paper is quite mathematical in nature, you should be able to explain it in your own words reasonably well without actually copying it.

  • Yes, I've actually already retype set the formula, with added working. for my benefit. But I was hoping there was away to improve the world by making this important paper easier to consume. Jan 18 '14 at 2:31
  • making a lecture note might be one way to do that.
    – Suresh
    Jan 18 '14 at 4:02
  • I guess so yes, and said lecture notes, I would be able to release. Now I am getting what you are saying! Jan 18 '14 at 4:09
  • 1
    "you can't reproduce figures/tables from the original work without permission": depending on your legislation this may be covered by a kind of fair use. E.g. in Germany if the purpose is teaching or research it would be permitted (of course properly citing the source). Jan 18 '14 at 17:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.